author manuscript nih public access norbou g. buchler leonard … · introduction post-mortem and...

Assessing the effects of age on long white matter tracts usingdiffusion tensor tractography

Simon W. Davis1,2, Nancy A. Dennis1,3, Norbou G. Buchler1,4, Leonard E. White3,5, DavidJ. Madden3,4, and Roberto Cabeza1,2,3,41 Center for Cognitive Neuroscience, Duke University, Durham, NC 27708, USA2 Department of Psychology and Neuroscience, Duke University, Durham, NC 27708, USA3 Center for the Study of Aging and Human Development, Duke University Medical Center, Durham,NC 277104 Brain Imaging and Analysis Center, Duke University Medical Center, Durham, NC 277105 Department of Community and Family Medicine, Duke University Medical Center, Durham, NC27710

AbstractAging is associated with significant white matter deterioration and this deterioration is assumed tobe at least partly a consequence of myelin degeneration. The present study investigated specificpredictions of the myelodegeneration hypothesis using diffusion tensor tractography. This techniquehas several advantages over other methods of assessing white matter architecture, including thepossibility of isolating individual white matter tracts and measuring effects along the whole extentof each tract. The study yielded three main findings. First, age-related white matter deficits increasedgradually from posterior to anterior segments within specific fiber tracts traversing frontal andparietal, but not temporal cortex. This pattern inverts the sequence of myelination during childhoodand early development observed in previous studies and lends support to a “last-in-first-out” theoryof the white matter health across the lifespan. Second, both the effects aging on white matter andtheir impact on cognitive performance were stronger for radial diffusivity (RD) than for axialdiffusivity (AD). Given that RD has previously been shown to be more sensitive to myelin integritythan AD, this second finding is also consistent with the myelodegeneration hypothesis. Finally, theeffects of aging on select white matter tracts were associated with age difference in specific cognitivefunctions. Specifically, FA in anterior tracts was shown to be primarily associated with executivetasks and FA in posterior tracts mainly associated with visual memory tasks. Furthermore, thesecorrelations were mirrored in RD, but not AD, suggesting that RD is more sensitive to age-relatedchanges in cognition. Taken together, the results help to clarify how age-related white matter declineimpairs cognitive performance.

KeywordsAging; DTI; tractography; radial diffusivity; demyelination

Corresponding Author: Simon W. Davis, Center for Cognitive Neuroscience, Duke University, Box 90999, LSRC Bldg., Durham, NC27708, USA, T: 919-668-2299, F: 919-681-0815, [email protected]'s Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customerswe are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resultingproof before it is published in its final citable form. Please note that during the production process errors may be discovered which couldaffect the content, and all legal disclaimers that apply to the journal pertain.

NIH Public AccessAuthor ManuscriptNeuroimage. Author manuscript; available in PMC 2010 June 1.

Published in final edited form as:Neuroimage. 2009 June ; 46(2): 530–541. doi:10.1016/j.neuroimage.2009.01.068.

NIH

-PA Author Manuscript

NIH


NIH


IntroductionPost-mortem and in vivo studies of the structural composition of the human brain converge onthe idea that aging is associated with significant changes in white matter architecture (Meier-Ruge et al., 1992; for review, see Raz, 2005; Resnick et al., 2003; Scahill et al., 2003; Tangand Nyengaard, 1997). Evidence from humans (Yakovlev and Lecours, 1967) and non-humanprimates (Peters et al., 2000) suggests that this deterioration is partly due to the degenerationof myelin in old age, which is a process likely to impair the conduction of neural signals acrossthe brain. The main goal of the current study was to investigate the possible predictions of thismyelodegeneration hypothesis using diffusion tensor imaging (DTI) tractography.

DTI tractography is a non-invasive imaging technique for investigating changes in white matterintegrity. Whereas early neuroimaging studies of white matter and aging typically usedqualitative measures of white-matter hyperintensities or quantitative analyses of white mattervolume (for review, see Sullivan and Pfefferbaum, 2006; Wozniak and Lim, 2006), more recentstudies have used DTI. This technique assesses the diffusion of water molecules, which is moreconstrained for intact than degraded white matter fibers (Concha et al., 2006; Thomalla et al.,2004). Most DTI studies of aging have focused on regions-of-interest (ROI) or clusters ofvoxels showing significant differences (e.g., Grieve et al., 2007; Madden et al., 2004). Thesemethods afford a number of benefits: manual ROI volumetry is the gold standard ofneuroanatomical assessment and affords a high degree of spatial precision when compared toother voxelwise methods (Kennedy et al., 2008); voxel-based morphometry (VBM; Ashburnerand Friston, 2000) approaches, one the other hand, benefit from a finer spatial localizability.The contributions of both ROI- and voxel-based analyses have been critical in identifying thefinding that age-related differences in white matter integrity are largely localized to the frontalcortex. However, these approaches are limited in their assessment of white matter anatomy inthat a single ROI or voxel cluster (1) includes several different white matter tracts connectingany number of different cortical areas, and (2) covers only a fraction of a particular long whitematter tract. Both of these limitations are addressed in diffusion tensor tractography becausethe tracts identified with this method can be more specific to an individual white matter tract(e.g., the uncinate fasciculus), and can assess the whole extent of a longitudinal tract. DTItractography therefore offers not only the opportunity to map fiber tracts representative of theunderlying anatomy (Burgel et al., 2006), but also allows for the ability to test specifichypotheses about age-related changes in white matter morphology along the length of specificfiber tracts. In the present study, we used a seed-based diffusion tensor tractography approach(Corouge et al., 2006; Mori and van Zijl, 2002) to investigate two predictions motivated bythe myelodegeneration hypothesis.

First, we investigated the prediction that age-related white matter deficits should increasegradually from posterior to anterior brain regions. Developmental studies have shown thatfrontal and temporal association regions are the last ones to myelinate (Kinney et al., 1988;Sakuma et al., 1991). Given the assumption that the brain regions that mature last duringdevelopment tend to decline first during aging—“last in, first out” (Raz, 2000)—one wouldexpect that age-related white matter should be more pronounced in anterior compared toposterior brain regions. DTI studies using ROI and VBM approaches have generally reportedgreater age effects in frontal compared to primary posterior cortices, though age-related declinein white matter volume and integrity occurs throughout the brain (Head et al., 2004; Maddenet al., 2007; Madden et al., 2006; O’Sullivan et al., 2001; Raz et al., 2005; Salat et al., 1999;Sullivan et al., 2006). While many of these previous studies have reported a general “anterior-posterior gradient” describing the pattern of age-related changes (Head et al., 2004; Maddenet al., in press; Pfefferbaum et al., 2005), this model has not been directly investigated in eithervoxelwise or ROI methods. We test one specific prediction of this model, in that the gradienthypothesis is specific to certain fiber systems within cerebral regions where the anterior-

Davis et al.

Neuroimage. Author manuscript; available in PMC 2010 June 1.

NIH


NIH


NIH


posterior gradient has been previously observed, namely within frontal and parietal whitematter. Tractography is well-suited to isolate these fiber systems and provides a means ofdirectly testing the question of whether the age-related differences found in these frontalregions reflects a general anterior-posterior gradient within specific tracts connecting anteriorand posterior regions, or instead reflects the selective vulnerability of frontal white matter.Given the debate about whether age-related cognitive decline is primarily due to frontaldysfunction (Moscovitch and Winocur, 1992; West, 1996) or to more widespread changes inthe brain (Greenwood, 2000), this is an issue with important theoretical implications. If theanterior-posterior gradient is in fact demonstrable in specific white matter tracts, age effectsshould increase gradually from posterior to anterior sections of each fiber tract traversing thefrontal lobe association regions. In contrast, if age-related decline is specific to frontal regions,age effects should increase abruptly at the point in which the fiber tract enters the frontal lobes.

Second, we investigated the prediction that the effects aging on white matter and their impacton cognitive performance should be stronger for radial diffusivity (RD) than for axial diffusivity(AD) measures. Evidence from post-mortem studies in primates suggests that the protractedage-related degeneration of white matter is at least partially related to changes in myelination,including increases in the number of dense inclusions, the formation of myelin balloons, or theformation of redundant myelin (Feldman and Peters, 1998; Peters and Sethares, 2002). Whilethese morphological changes may underlie the observed age-related differences in white mattervolume or integrity, in vivo evidence for myelodegeneration in humans is limited. Althoughmost DTI studies have focused on fractional anisotropy (FA) as a summary measure of whitematter integrity, there is evidence that two components of the diffusion signal, namely AD andRD, have different relations to cellular mechanisms of white matter deterioration. AD describesthe principal eigenvector (λ1) and is assumed to contribute information regarding the integrityof axons (Glenn et al., 2003) or changes in extra-axonal/extracellular space (Beaulieu andAllen, 1994). In contrast, RD describes an average of the eigenvectors perpendicular to theprincipal direction ( [λ2 + λ3]/2 ) and is assumed to characterize changes associated withmyelination or glial cell morphology (Song et al., 2003; Song et al., 2002; Song et al., 2005).This relationship has recently been extended to humans in a study by Schmierer and colleagues(2004), in which a combined diffusion imaging and histological assessment of unfixed braintissue collected post mortem from multiple sclerosis (MS) patients showed that both FA andRD, but not AD, were significant predictors of myelin content. While this direct relationshipbetween RD and myelin has yet to be extended to healthy adults, DTI studies in healthy olderadult populations have generally found greater age-related declines in the medium and minoreigenvectors, indicating a decline in radial diffusivity (Bhagat and Beaulieu, 2004; Madden etal., in press; Sullivan et al., 2006; Sullivan et al., 2008; Vernooij et al., 2008). Furthermore,given that studies have shown that developmental maturation of white matter may primarilybe driven by decreases in RD (Giorgio et al., 2008; Snook et al., 2005), it is reasonable tohypothesize that changes to white matter anatomy later in life are also a consequence ofmyelination, and maybe evidenced by increases in RD in fiber systems that are the latest tomyelinate (Lebel et al., 2008). Thus, the myelodegeneration hypothesis predicts that age effectsshould be larger for RD and AD, and that these effects should be observed in all late-myelinating regions, including both frontal and temporal association cortices. Moreover, thishypothesis predicts that correlations between white matter integrity and cognitive performancein older adults should be stronger for RD than for AD measures.

In addition to testing the foregoing two predictions derived from the myelodegenerationhypothesis, a third goal of the study was to investigate the specificity of associations betweenthe effects of aging on different white matter tracts and different cognitive measures. Giventhat (a) each cognitive function depends on a certain set of brain regions and correspondingwhite matter connections and (b) that the effects of aging across the brain show substantialindividual differences (Glisky et al., 1995; Glisky et al., 2001), it is reasonable to expect that

Davis et al.


NIH


NIH


NIH


correlations between the effects of aging on white matter tracts and cognitive tasks should bespecific rather than global. Examining this idea requires assessment of several different tractsand several different cognitive measures, but most previous DTI studies of aging have focusedon a few white matter ROIs and a few cognitive measures (Bucur et al., 2007; Madden et al.,2004; Persson et al., 2006; Sullivan et al., 2001; Sullivan et al., 2006). Two studies investigatedDTI-behavior correlations using larger set of cognitive tests (Grieve et al., 2007; O’Sullivanet al., 2001), but their conclusions regarding specificity are limited because they foundsignificant age effects only in executive functions tests. To address this issue, we investigatedcognitive performance using a standardized neuropsychological battery (Owen et al., 1996;Robbins et al., 1998) likely to yield reliable age effects on several different tasks, includingtests of memory as well as executive functioning. We then correlated DTI measures andcognitive performance measures to identify tracts that were significant predictors of age-relateddifferences in cognitive performance in older adults.

MethodsParticipants

The participants were 20 older adults (12 female; M = 68.89 years, SD = 5.3 years,) and 20younger adults (9 female; M = 20.04 years, SD = 2.5 years) with normal or corrected-to-normalvision, and no history of neurological or psychiatric disease. Written informed consent wasobtained from each participant, and the study was approved by the Duke UniversityInstitutional Review Board. Younger adults were recruited from the Duke Universitycommunity; older adults were community-dwelling individuals who were recruited from thegreater Durham community. Older adults were screened for health problems and conditionsthat could affect blood flow (e.g., hypertension, medications affecting blood flow) using aquestionnaire. The selected group of older participants performed well within the normal rangein the Mini-mental Status Exam, as well as the normal age appropriate range on tests ofcognitive function included in the Cambridge Neuropsychological Test Automated Battery(CANTAB), ensuring that they were cognitively healthy and free from dementia, with allimaging performed within 5 months of testing.

Behavioral TestingBefore scanning, participants in both the older and younger groups underwentneuropsychological assessment with the CANTAB battery (Table 1). Included were four testsassociated with executive functioning (Intra-Extra Dimensional Set Shifting, Spatial Span,Spatial Working Memory, Rapid Visual Information Processing), two tests of visual memoryperformance (Paired-Associate Learning, Pattern Recognition Memory), and two processingspeed tasks (Movement Time and Reaction Time).

DTI Acquisition and PreprocessingMRI data were acquired on a 3T General Electric (Milwaukee, WI) Signa human MRI scanner.The DT-MRI dataset was based on a single-shot EPI sequence (TR = 1700 ms, 50 contiguousslices of 2.0 mm thickness, FoV = 256 × 256 mm2, matrix size 128 × 128, voxel size 2 × 2 ×2 mm, b-value = 1000 s/mm2, 15 diffusion-sensitizing directions, 960 total images, total scantime approx. 5 min). The resulting images were skull-stripped in order to remove non-brainand ambient noise using the Brain Extraction Tool (Smith et al., 2002), and eddy-corrected toaccount for drifts in scanner acquisition. Diffusion tensors were calculated from the 15diffusion weighted images based upon a typical simple least squares fit of the tensor model tothe diffusion data (Basser et al., 2000). Diagonalization of the tensor yielded three voxel-specific eigenvalues (λ1 > λ2 > λ3), which represent diffusivities along the three principaldirections of the tensor. These three eigenvalues were used to construct fiber tracts and thediffusion properties (RD, AD) discussed below.

Davis et al.


NIH


NIH


NIH


Fiber TrackingDiffusion fiber tracts were constructed using software distributed by Gerig and colleagues(http://www.ia.unc.edu/dev/), based on the method of Mori and colleagues (Mori and van Zijl,2002; Xu et al., 2002), which applies a unit-step algorithm to tensor streamlines following thedirection of greatest diffusivity (λ1) between seed and target ROIs (for a more detailedexplanation, see Catani et al., 2002). The method consisted of four steps (see Fig. 1). First,seed and target ROIs for each tract of interest were drawn on the native space b0 (non-diffusion-weighted) image and fibers traversing seed and target regions were estimated by the software(Fig. 1A). Consistent with previous tractography studies using similar methods (Sullivan etal., 2006; Wakana et al., 2007), fiber tracking parameters included a minimum FA of 0.2, afiber tracking threshold of 0.125 and an angle of maximum deviation of 35° Additionalparameters including seed and target ROIs and minimum/maximum fiber length weredetermined specific to local fiber tract characteristics and standard anatomical landmarks; theseare described for each fiber tract in Table 2. Fibers were then represented as polylines andreparameterized by cubic B-spline curves (Fig. 1B), which ensured an equidistant samplingalong each fiber as well as consistent sampling for all fibers (Corouge et al., 2006). Third, afterfiber estimation, mean tensor fibers were constructed from each set of fibers (Fig. 1C).Diffusion properties were then computed and reparametrized according to the distance froman established origin based upon a consistent anatomical landmark; diffusion properties wereevaluated at points along the length of each fiber at cross sections. Critically, all fiber tractsremained in native space throughout fiber tracking procedure, and only after parameters werederived from the individual were values adjusted for normalized fiber length and averagedacross age group. Fourth, after additional processing and statistical analysis (see below), groupdifferences were transposed onto the mean tract for visualization (Fig. 1D).

The resulting parameters included mean fractional anisotropy, axial diffusivity (λ1), and radialdiffusivity ([λ2 +λ3]/2). To derive diffusion properties evident of consistent locations acrosssubjects, we set anchor points at consistent anatomical loci for each tract of interest locatedalong the path of the fiber tract (See Table 2). In order to ensure sampling from a uniformanatomical structure and not individual fibers deviating from the main axis of the tract, meanfibers were computed for each tract in each subject by averaging the spatial characteristics ofall extracted polylines in a fiber tract, along an axis extending from the tract midline. Individualtractography fibers deviating more than 35° from the mean fiber were then excluded fromanalysis. In order to account for individual variability in fiber lengths, tract data were thenlinearly adjusted across subjects via linear interpolation according to at least 3 anatomicallandmarks along the course of the fiber before group averages were calculated, and thensectioned into 10mm segments along the main axis of the tract.

Anatomical guidelines and termination heuristics for defining tracts of interest were determinedby anatomical reference (Bassett, 1952; Curran, 1909; Duvernoy, 1999; Ebeling and vonCramon, 1992; Kier et al., 2004), and a diffusion tractography atlas (Wakana et al., 2004) andwere supervised by an expert neuroanatomist (LEW). Seed and target ROIs for tractographyare represented in Table 2. Additionally, our fiber tracking was performed by 2 raters; inter-rater reliability was assessed by performing repeated fiber tracking measurements on the samesubset of 6 individuals randomly assigned for gender and group for each tract and raters wereblind to age group of the data, such that tractography procedures did not differ between groups.Briefly, fiber tracts were first converted to binary volumetric information by assigning a valueof 1 or 0, and difference images between raters were used to assess the spatial overlap in pixelsbetween two raters. The Cohen’s kappa coefficient was calculated by K = (observed agreement− expected agreement)/(1 − expected agreement). The analysis was applied pair-wise betweenindividual raters, providing an intra-rater reliability estimate for each tract (see Table 2).Previous studies using similar methods have described the strong intra-rater reliability of this

Davis et al.


NIH


NIH


NIH


http://www.ia.unc.edu/dev/

approach (Wakana et al., 2007), indicating that the current methods provide a reliable estimateof between expert raters. All tracts met the fiber tracking parameters described above, exceptfor one younger adult right UF and one older adult left UF, likely due to scanner artifact in thissusceptible inferofrontal region. The remaining white matter structures traced in this study allshowed high inter-rater reliability across all tracts, with average IRR scores within the “almostperfect” agreement (K > 0.80) set forth by Landis and Koch (1977). Additionally, we comparedboth the total number of extracted fibers as well as number of outliers (# of fibers before outlierremoval - # fibers after outlier removal) on each fiber tract and found no significant effects ofage (p < 0.05); there were no effects of gender on FA, RD, or AD across all tracts, which isconsistent with previous findings (Salat et al., 1999; Sullivan et al., 2001).

Supplementary Voxelwise ValidationIn order to validate our tractography results a follow-up voxelwise analysis was performedusing Tract-Based Spatial Statistics (TBSS; Smith et al., 2006) module of FSL (Smith et al.,2004). TBSS is an attractive means of comparing tractography with voxel-based methods, inthat the normalization used is particularly responsive to the underlying white matterarchitecture. In accordance with the standard TBSS pipeline, FA maps from all 40 subjectswere normalized to the FMRIB58_FA standard space white matter template. A voxelwiseanalysis of the normalized FA maps was then completed in a two-sample (young, old) t-test inStatistical Parametric Mapping 5 (SPM5; http://www.fil.ion.ucl.ac.uk/spm/). Given that tractswithin the current analysis observed an anterior-posterior gradient only when traversingfrontal-parietal cortices (see below), we applied both a whole-brain WM as well as a frontal-parietal WM mask to the voxelwise analysis above. The whole brain mask comprised all whitematter voxels in the FMRIB58_FA atlas with an FA value greater than 0.2; the frontal-parietalmask comprised a subset of those voxels located within the frontal and parietal lobe ROIs fromthe PickAtlas toolbox (http://www.fmri.wfubmc.edu/download.htm) in SPM5.

Statistical analysesTo investigate the prediction that age-related white matter deficits should gradually increasefrom posterior to anterior brain regions, we focused on the effects of aging along the y-axis oftwo tracts that follow an anterior-posterior trajectory and enter the frontal lobe: the cingulumbundle (CB) and frontal section of the uncinate fasciculus (UF). Temporal sections of this lattertract were excluded from trend analysis by removing points inferior to the fronto-temporalborder, which also served as the midpoint landmark (i.e., anchor) for this fiber tract (see Table2). Adjusted tract coordinates from the group means above were transposed linearly toTalairach coordinates, and lengths of each tract were divided into 10mm sections with respectto their position along the y-axis. To distinguish between gradual vs. abrupt changes in ageeffects, independent-samples t-test was performed at each point along a normalized fiber tract;trend analyses were performed on the resulting measure of age differences in FA (partial eta)along the anteroposterior extent of the fiber tract. Both linear and quadratic changes were testedsimultaneously in the model. In this analysis, a linear trend represents a gradual change, aquadratic trend represents an abrupt change, and the intercept represents the main effect of age.Thus, a trend analysis can be useful in characterizing whether age-related differences along atracts are global (intercept differences) or whether they differ along the tract gradually (linearterm) or abruptly (quadratic term) along the anteroposterior extent of a longitudinal fiber tract.Trend analyses were repeated on t-statistics from the masked voxelwise analysis describedabove, averaging the t-statistic for each slice within the mask along the same anterior-posterioraxis as in the above tractography-based analysis (see Figure S1).

To investigate the prediction that the effects aging on white matter and their impact on cognitiveperformance should be stronger for radial diffusivity (RD) than for axial diffusivity (AD)measures were performed two analyses. First, we performed age (younger vs. older) x

Davis et al.


NIH


NIH


NIH


http://www.fil.ion.ucl.ac.uk/spm/

http://www.fmri.wfubmc.edu/download.htm

diffusivity measure (RD vs. AD) ANOVAs on all sections of all tracts showing significant ageeffects on FA. Age group and diffusivity measures were used as fixed factors, subjects as arandom factor, and gender as a covariate of interest. Additionally, RD and AD were evaluatedindependently for group effects, with gender again included as a covariate of interest. Second,using t-tests we tested whether the correlations between age-related white matter decline andage-related cognitive decline (see below) were greater when for RD than AD.

Finally, to investigate the specificity of associations between the effects of aging on differentwhite matter tracts and different cognitive measures we correlated DTI measures and CANTABmeasures. The goal of these analyses was to find tracts which act as predictors of cognitiveperformance in older adults; hence, only test scores and white matter tracts which showed botha significant age-related difference in FA (p < 0.01) and test performance (p < 0.05) wereincluded in the analysis. Because age is positively correlated with white matter decline, weused a stepwise regression model in order to identify age-dependent changes in white matter-behavior relationships. The Age variable was forced into the regression model at the first step,and other predictor variables were entered into the model sequentially based on the strengthof their relationship to the dependent variable, covaried for the effects of variables enteredpreviously. Model estimation ended when no other variable met the criterion for entry (p =0.5). Significant Age x FA interactions predicting task performance were identified and furtherevaluated for significant effects within older adults with Pearson’s r.

ResultsAnterior-posterior trend analysis

Figure 2 shows age-related differences in FA (younger – older) along the length of five tracts:the genu and the splenium of the corpus callosum, the cingulum bundle (CB), the inferiorlongitudinal fasciculus (ILF), and the uncinate fasciculus (UF). Qualitative inspection of thisfigure suggests that the greatest age-related differences in FA (i.e., warmer colors in Figure 2)occurred mainly in anterior segments of longitudinal tracts traversing the frontal lobe. Toinvestigate the prediction that age-related white matter deficits should gradually increase fromposterior to anterior brain regions, we focused specifically on two tracts that follow an anterior-posterior trajectory and enter into the frontal lobes, the UF and CB. As illustrated by the trendanalyses in Figure 4, in both left and right UF and in both left and right CB, group-related effectsizes from an Young > Older t-contrast increased linearly from posterior to anterior regions.To determine if the change was gradual or abrupt, linear and quadratic regressors weresimultaneously tested. The intercept trend (p < 0.01) was significant for tracts, indicating amain effect of age. Importantly, the linear regressor was significant in all four tracts (left andright frontal UF, left and right CB), whereas the quadratic regressor was significant for noneof them. Follow-up analyses based upon a masked voxelwise analysis confirmed these lineartrends for voxels within frontal and parietal cortex (linear: t = 5.27, p < 0.01; quadratic: t =1.13, p = n.s.), providing an important validation of the tractography results compared to morestandard voxelwise analyses. Combined, these results support the prediction of themyelodegeneration hypothesis that age effects increase gradually from posterior to anteriorregions rather than being specific to the frontal lobes.

Radial vs. axial diffusivityAge-related differences in RD and AD for all tracts are reported in Table 3 and depicted inFigures 2 and 3. To investigate the prediction thatthe effects of aging are greater for RD forAD, we tested for Age (younger, older) x Diffusivity Measure (RD, AD) interactions.Significant interactions were found in the genu (F = 11.39; p < 0.005), bilateral CB (L: F =6.39; p < 0.05; R: F = 5.90; p < 0.05) and bilateral UF (L: F = 5.69; p < 0.05; R: F = 5.49; p <0.05), and in all cases age differences were greater for radial than axial diffusivity (Figure 3).

Davis et al.


NIH


NIH


NIH


In left and right ILF, the interaction did not reach significance, but the age effect was significantfor RD but not for AD, consistent with the pattern displayed by the other tracts in late-myelinating regions.

Tract-Function SpecificityTo investigate specific relations between the effects of aging on particular white matter tractand particular cognitive functions we focused fiber tracts that showed significant age-relateddecline in FA (see Table 3) and cognitive tasks that showed reliable age effects (see Table 1).Within older adults, a number of correlations between task performance and diffusion metricswere found and are displayed in Table 4 and Figure 5. As evidenced by the standard deviationsin Table 1, data from younger adults’ performance showed a relatively low variance comparedto older adults; consistent with other tractography studies of younger adults, diffusivitymeasures for these subjects also showed a similarly low variance in FA, AD, and RD values(see Table 3). It is therefore unsurprising that no significant relationship was observed betweenbehavior and structure in this younger population. Consistent with the notion of functionalspecificity, tracts connecting frontal regions in older adults (e.g., genu, right UF) were highlycorrelated with performance on executive function tasks (Spatial Span, Spatial WorkingMemory, and Intra-Extra Dimensional Set Shifting), whereas tracts connecting more posteriorregions (left CB, ILF, splenium) were highly correlated with performance on visual memorytests (Paired Associate Learning, and Pattern Recognition Memory).

We also investigated how well RD and AD predict cognitive deficits associated with aging.As illustrated by Table 4, correlations were generally significant for RD but not for AD.Importantly, with the exception of two posterior tracts (left ILF and splenium), the correlationswere significantly stronger for RD than for AD; t-tests showed greater behavior-diffusivitymeasure correlations for RD than AD in the genu (Spatial Span: t = 5.90, p < 0.01), right UF(Spatial Working Memory: t = 3.50, p < 0.01; Spatial Working Memory-strategy: t = 2.47, p< 0.05; Intra-Extra Dimensional set shifting: t = 3.97, p < 0.01) and left CB (Pattern RecognitionMemory: t = 4.89, p < 0.01; Paired Associate Learning: t = 2.48, p < 0.05). In sum, the resultssuggests that the effect of aging on individual white matter tracts influences specific cognitivefunctions, and support the prediction of the myelodegeneration hypothesis that the effects agingon white matter and their impact on cognitive performance are stronger for RD than AD.

DiscussionOur results revealed three main findings. First, age-related white matter deficits within tractstraversing frontal and parietal cortices increased gradually, not abruptly, from posterior toanterior brain regions. Second, the effects of aging on white matter structure and their impacton cognitive performance were stronger for RD than for AD measures. Finally, results revealedthat the effects of aging on particular white matter tracts had an impact on specific cognitivefunctions; in particular, anterior tracts were primarily associated with executive tasks andposterior tracts were mainly associated with visual memory tasks. Furthermore, the resultsshow that relationships between cognitive performance and FA were generally mirrored in RD,but not AD. These findings support predictions based on the myelodegeneration hypothesis ofaging and are discussed in separate sections below.

Anterior-Posterior GradientThe first main finding of the study was that age-related white matter deficits increased graduallyfrom posterior to anterior brain regions within specific white matter tracts. The anterior-posterior gradient was found within tracts that traverse frontal and parietal cortex (CB & UF),but not for tracts which traverse the temporal lobe (ILF). Thus, the data partially supports thegeneral anterior-posterior gradient hypothesis in that it may be generalized to only those fiber

Davis et al.


NIH


NIH


NIH


systems within frontal and parietal lobes. This new finding of a linear gradient of change withinfiber tracts traversing through and outside the frontal cortex does not support the frontalhypothesis of aging. Although this finding is consistent with the results of ROI-based DTIstudies that have identified an increase in the age-related difference in FA from posterior toanterior regions (Bucur et al., 2007; Head et al., 2004; Salat et al., 2005), it adds to these findingsby providing greater specificity to the previously described anterior-posterior gradient (Headet al., 2004; Madden et al., in press; Pfefferbaum et al., 2005) in specific tracts connecting thefrontal lobe with more posterior cortices. As noted in the Introduction, ROI-based DTI analysescannot distinguish among overlapping white matter fiber tracts within a given ROI, and do notnormally sample the whole extent of long white matter fibers. In contrast, the tractography-based analyses employed in this study were able to isolate individual fiber tracts highlyrepresentative of white matter anatomy, and measure age effects along the whole extent of eachtract.

Measuring age effects along the whole extent of each tract is particularly important for the goalof investigating anterior-posterior differences in age-related white matter decline because itallows one to (1) compare age effects on anterior vs. posterior sections of the same tract, asopposed to comparing different age effects on separate anterior vs. posterior tracts (e.g., genuvs. splenium), and (2) measure age effects continuously along each fiber tract. Because of thesefeatures, the current study was able to compare two possible patterns of anterior-posteriordifferences in age-related white matter decline: a gradual pattern vs. frontal-specific pattern.As illustrated by Figure 4, significant linear trends supported the idea of a gradual pattern,though only within specific white matter fiber tracts. In the cingulum bundle (CB), for example,age effects were greater for frontal than for posterior brain regions, and they decreasedgradually from anterior frontal (e.g., y = 20) to posterior frontal (e.g., y = 0) to anterior parietal(e.g., y = −20) and to posterior parietal (y = −40) regions. This finding challenges the traditionalidea that healthy cognitive aging is a frontal phenomenon (Moscovitch and Winocur,1992;West, 1996) and suggest instead that age effects decrease gradually from anterior toposterior brain regions.

Importantly, these linear trends identified in our tractography analysis were replicated in amasked voxelwise analysis, in that an anterior-posterior gradient also described age-relateddifferences throughout the frontal and parietal lobes. Thus, results of both tractography andvoxelwise assessments provide greater specificity to the anterior-posterior gradient, localizingthis effect to white matter connecting frontal and parietal cortices. Additionally, bothtractography and voxelwise analyses (see Figure 2, S1) identified an exception to the anterior-posterior trend within temporal lobes (discussed below), suggesting that this trend does notgeneralize to temporal cortices. While future studies are needed to further evaluate thecorrespondence between analysis techniques, the identification of a linear trend within thevoxel-based analysis nonetheless suggests further support for the anterior-posterior gradientwithin frontal and parietal cortices.

As noted in the Introduction, one possible explanation of the observed anterior-posteriorgradient is the “first-in-last-out” hypothesis (Raz, 2000), which postulates that white matterdecline during aging follows the reverse order of white matter maturation during childhooddevelopment. Evidence from developmental studies suggests that during childhood thematuration of both gray and white matter roughly follow a posterior-to-anterior gradient, withmore anterior/frontal regions developing last (Giedd et al., 1999; Sowell et al., 2002).Furthermore, it is well established that during cortical maturation, white matter within posteriorsensory areas develop first whereas anterior association areas develop at a later stage(Eluvathingal et al., 2007; Flechsig, 1901; Paus et al., 1999; Sowell et al., 2004; Yakovlev andLecours, 1967). For example, primary visual cortex is one of the first regions to develop fullyformed myelin sheaths, whereas anterior association regions begin myelination much later in

Davis et al.


NIH


NIH


NIH


development and continue to undergo myelination until the second or third decade of life(Hayakawa et al., 1991; Kinney et al., 1988). Additionally, the “last-in-first-out” hypothesis isalso reflected by a rank ordering based on effect sizes for each tract for the age difference inFA (genu > R UF > L UF > R ILF > L ILF > R CB > L CB > splenium), and suggests the samegeneral pattern consistent with the last-in, first-out hypothesis: tracts showing the largest effectsin the current study are the same tracts that tend to develop later in the first decade of life(Dubois et al., 2008; Eluvathingal et al., 2007). For example, the occipito-temporal segmentof the ILF showed a greater age effect in FA than more anterior regions of the fiber tract; theboundary between occipital and temporal cortices is a late-myelinating region (Flechsig,1901), and hence, the first-in-last-out hypothesis—but not the gradient hypothesis—wouldpredict early age-related white matter decline in this region.

The effects of age on white matter architecture are complex and cannot be completely explainedby the anterior-posterior gradient. Notably, (1) not all fiber systems observe this pattern, and(2) age-related differences in RD (Figure 3) did not appear to observe the same anterior-posterior pattern as observed in FA (Figure 2). One explanation for this discrepancy may bethat FA describes only a general pattern of age-related changes in white matter, and taken asindividual components AD and RD may represent qualitatively different patterns not dependenton position along an anterior-posterior axis. A more precise description of age-related changenecessitates both a greater spatial specificity to the previously described anterior-posteriorgradient, as well as complimentary metrics of white matter health such as RD and AD.Therefore, a description of the whole-brain pattern of age-related changes may not be bestdescribed by the anteroposterior extent but instead by the developmental history of the regionthe tract traverses.

Radial vs. Axial DiffusivityThe second main finding of the study was the effects aging on white matter and their impactson cognitive performance were stronger for RD than for AD measures (See Table 3 and Figures2 and 3). Taken together with previous neurobiological studies (Song et al., 2005;Sun et al.,2007;Trip et al., 2006) and complementary findings in callosectomy patients (Concha et al.,2006) and MS (Schmierer et al., 2004) suggesting dissociations between RD/myelination andAD/axonal integrity, this finding suggests that age-related differences in FA are more likelydriven by myelodegeneration than by a loss of axonal integrity. When compared directly, theage-related changes associated with AD to that of RD in the genu, CB (left and right), and UF(left and right) all showed significant Age x Diffusivity Measure interactions. Moreover, nowhite matter regions investigated herein showed significant differences in AD within theconservative constraints of the present study (p < 0.01). Thus, the current finding lendsconsiderable support to the idea that age-related changes in white matter are largely driven bychanges in myelination rather than by gross morphological changes in the axons themselves(Bartzokis, 2004). It should be noted that increases in RD likely reflects other systematic age-related changes in the characteristics of white matter tracts aside from demyelination, includingreductions in the concentration of neuroglia and changes associated with gliosis (Beaulieu andAllen, 1994). While there has been no direct link between diffusivity measures as recordedfrom DTI and post mortem analysis of myelin content in healthy older adult humans, theobservation from previous studies that oligodendrocytes and other myelin-supporting glia areparticularly susceptible to the effects of aging (Bartzokis et al., 2004;Peters and Sethares,2002) nonetheless suggests a link between RD and the health of myelin in the aging brain.Furthermore, a recent analysis of post mortem MS brain tissue has shown FA and RD—butnot AD—to be significant predictors of myelin content (Schmierer et al., 2004), providingsome of the most compelling support for the relationship between RD and myelin content.Future post mortem studies investigating the relationship between diffusivity measures and

Davis et al.


NIH


NIH


NIH


histological markers of myelin content (e.g., via myelin staining) within brain tissue fromhealthy older adults are necessary to provide further support for this hypothesis.

Previously, we discussed how the observed age differences in white matter integrity followedan anterior-posterior gradient and that this gradient could, in turn, be explained by last-in-first-out hypothesis of development and aging. The current pattern of age-related differences in RDcan also be readily understood within this last-in-first-out framework. For example,developmental studies have shown that white matter near cortical association areas that developlater in maturation (Lebel et al., 2008) also shows the earliest age-related decline in myelination(Bartzokis, 2004; Raz et al., 1997; Yakovlev and Lecours, 1967). This idea is supported in thecurrent results, with fibers near frontal and temporal association cortex showing the largestgroup differences in RD (See Table 3). Moreover, a rank ordering of the effect size in the groupcomparison of RD—(R CB > L ILF > L CB > R ILF > genu > R UF > L UF > splenium) makesclear the observation that this pattern of age-related decline may at least partially invert thesequence of myelination observed during childhood and development. While studies usinglarger age ranges will be necessary to test this idea directly, the current findings demonstratesthe differences between generalized patterns of white matter decline (i.e., the anterior-posteriorgradient), and patterns that may be more representative of the degree of myelination.

These results compliment recent DTI studies of infants (Gao et al., 2008) and young adults(Lebel et al., 2008) that demonstrate that many of the same frontal fiber systems (e.g., UF, CB)mature later in development than more posterior fiber systems (e.g., ILF). A similar patternwas shown within interhemispheric fiber systems, with the splenium finishing majordevelopment before the genu. Later-myelinating regions have been shown to have a greaterprevalence of smaller axons, which have thinner myelin sheaths (Lamantia and Rakic, 1990)making them more vulnerable to the effects of age. Oligodendrocytes in these later regionshave also been shown to support a greater number of myelin sheaths (~50) thanoligodendrocytes in earlier-myelinating regions (10 or less), which may make those regionsmore susceptible to oxidation (Husain and Juurlink, 1995) and amyloid beta-proteinneurotoxicity (Geula et al., 1998). While more work is necessary to provide a comprehensivein vivo analysis of these effects, the current results provide strong support for a primary myelin-dependent mechanism underlying this pattern of age-related changes in white matter.

In addition to the general effect of aging on RD vs. AD, results found RD to be a better predictorcognitive performance in older adults. In fact, most of the diffusivity-behavior correlationsevaluated in the current study were significantly greater for RD than AD. This finding provideseven further support for the myelodegenerative hypotheses of aging, in that the same behavioralcorrelations between performance and general white matter integrity (i.e., FA) were mirroredin correlations between performance and a measure of myelination (i.e., RD). While it isrelatively well established that white matter plays a critical role in mediating cognitiveperformance (Wozniak and Lim, 2006), the current findings link these behaviors with a specificcomponent of the diffusion imaging signal (RD) and the cellular structure it may represent(myelin). While further research is needed to validate these in vivo assessments of myelination(e.g., through the complementary use of other imaging methods such as magnetization transferimaging), the current results offer strong support for the myelodegeneration hypothesis of aging(Bartzokis, 2004) and suggest a prominent role for this metric in future analyses of lifespandata.

Tract-Function SpecificityFinally, the study found that the effects of aging on particular white matter tracts had an impacton specific cognitive functions, with anterior tracts primarily associated with executive tasksand posterior tracts mainly associated with visual memory tasks. Within anterior tracts, FAand RD within the genu correlated significantly with Spatial Span, a measure of working

Davis et al.


NIH


NIH


NIH


memory ability. The fact that this correlation was observed only in older adults suggests thecross-hemispheric communication engendered by this tract may be important for maintainingcognitive performance in these subjects (Cabeza, 2002; Daselaar and Cabeza, 2005; Sullivanet al., 2006). Additionally, the right UF, which connects anterior temporal and infero-frontalregions, showed significant correlations with Spatial Working Memory and Intra-ExtraDimensional Set Shifting tasks, both measures of executive functioning. This structure likelyplays a significant role in mediating executive and memory functions (Markowitsch, 1995),and recent DTI tractography studies have shown it to be degraded in neuropsychologicalpatients who exhibit executive impairments (Price et al., 2008; Yasmin et al., 2008). In accordwith previous functional imaging studies, the results thus demonstrate the importance offrontotemporal connectivity for successful performance in healthy aging (Daselaar et al.,2006; e.g., Grady et al., 1995; Gutchess et al., 2005).

Turning to tracts connecting posterior regions of the cerebrum, the CB, ILF, and spleniumshowed significant correlations with scores on the Paired Associate Learning and PatternRecognition Memory tests. These tests use abstract figures to evaluate visual memoryperformance, which is functionally supported by the temporal, parietal and occipital regionsconnected by these tracts (Gould et al., 2003). Previous studies have shown paired associatelearning to show significant age-related decline (Rabbitt and Lowe, 2000), and it is wellestablished that older adults have more general impairments in visuospatial processing(Madden et al., 2006; Madden et al., 2002). The left CB showed a significant relationship withvisual memory performance, and this finding provides further evidence that this white matterstructure, and the cingulate cortex it traverses, subserve evaluative functions, such asmonitoring sensory events in the service of spatial orientation and memory (Vogt et al.,1992).

All told, the observed relationships between white matter structure and cognitive performanceare consistent with previous DTI studies using both younger and older adults. The presentfindings expand upon previous work by demonstrating that observed morphological changesattributable to distinct components of the diffusion signal may be an important mechanism ofage-related cognitive decline. The results further suggest that the effects of aging on individualwhite matter tracts influence specific cognitive functions reliant on the distributed neuralsystems underlying successful performance.

Limitations of the Current StudyThough the method presented here represents a novel algorithm for assessing the effects ofaging on long white mater association fibers, the findings from the current study must bequalified with some caveats. Although the pattern of age-related differences in the CB and UFshow that both tracts demonstrate a linear gradient of change, the extent to which this trendextends to other fiber systems may be dependent on local anatomy, especially for tracts whichdo not observe an anterior-posterior trajectory. As noted above, the progression of white matterdevelopment over the lifespan is complex and cannot be fully characterized by a linear gradient.We show here neuroimaging evidence that the observed trends may be a consequence of amore specific underlying pathology, such as myelin degeneration. However, the inferencesdrawn from neuroimaging data largely rely on previous research, and it is possible that theobserved changes in RD may be due to changes not in myelin but instead to the extracellularmatrix within the white matter; DTI alone cannot differentiate with certainty betweenintracellular and extracellular morphological changes in brain tissue, nor between effects toneurons or to glia. It is likely that combinations of newer imaging methods will likely providea more definitive explanation to the mechanism underling age-related changes in white matterin vivo. While previous methodological studies of white matter have demonstrated that DTItractography provides fiber tracts that have been shown to be highly representative of the

Davis et al.


NIH


NIH


NIH


underlying white matter anatomy (Burgel et al., 2006; Wakana et al., 2007), newer samplingmethods which are less susceptible to partial volume effects (Jensen et al., 2005) may improvethe inferences made on DTI imaging data. More practically, current methods of characterizingwell-known fiber systems using seed-based tractography are typically time-consuming and aresusceptible to human errors during tract tracing; more recent techniques relying on automatedclassification algorithms of these fiber systems (O’Donnell and Westin, 2006) will likelyimprove the efficacy of this technique. As noted above, more work is necessary to describe thecorrespondence between deterministic (e.g., seed-based tractography) and probabilistic (e.g.,TBSS) methods, as well as how well these methods characterize the complex architecture ofwhite matter (Goodlett et al., 2008; Ronen et al., 2005). The success of future age-relatedstudies of white matter structure will likely rely on the ability of tractography to complimentvoxel-based analyses, and the extent to which these methods can dissociate tract-specific fromregionally-specific effects. The present analysis nonetheless adds contribution to the currentliterature by carefully examining diffusion characteristics not only within particular fibersystems, but along the spatial trajectory of those fiber systems.

ConclusionsIn summary, the DTI tractography findings presented here make three significant contributionsto the literature characterizing age-related changes in white matter integrity. First, DTItractography results support an anterior-posterior gradient in age-related white matterdegradation within specific long-range white matter tracts traversing frontal and parietal cortex.This finding demonstrates a gradual age-related decline in white matter integrity along theanteroposterior extent of a tract, rather than a localized frontal degradation. Second, the age-related deficits in white matter integrity and their impact on cognitive performance were greaterfor RD than for AD, suggesting that the observed trends may reflect demyelination. Finally,cognitive performance in older adults was correlated with the integrity of specific white mattertracts throughout the cerebrum: frontal tracts were found to predict executive or workingmemory functions and more posterior white matter tracts were found to predict visual memoryperformance. Taken together the results characterize age-related changes in white matterintegrity and their role in age-related cognitive decline.

Supplementary MaterialRefer to Web version on PubMed Central for supplementary material.

AcknowledgmentsThis work was supported by a NIH grants AG19731 to RC and AG011622 to DJM. NAD was supported by NIA grantT32 AG000029. The authors would like to thank Amber Baptiste-Tarter forassistance in participant recruitment, JaredStokes, Vanessa A. Thomas, Matthew Emery and Jamaur Bronner for assistance with data collection, and Josh Bizzelland James Kragel foranalysis support.

ReferencesAshburner J, Friston KJ. Voxel-based morphometry--the methods. Neuroimage 2000;11:805–821.

[PubMed: 10860804]Bartzokis G. Age-related myelin breakdown: a developmental model of cognitive decline and

Alzheimer’s disease. Neurobiol Aging 2004;25:5–18. [PubMed: 14675724]author reply 49–62Bartzokis G, Sultzer D, Lu PH, Nuechterlein KH, Mintz J, Cummings JL. Heterogeneous age-related

breakdown of white matter structural integrity: implications for cortical “disconnection” in aging andAlzheimer’s disease. Neurobiol Aging 2004;25:843–851. [PubMed: 15212838]

Basser PJ, Pajevic S, Pierpaoli C, Duda J, Aldroubi A. In vivo fiber tractography using DT-MRI data.Magn Reson Med 2000;44:625–632. [PubMed: 11025519]

Davis et al.


NIH


NIH


NIH


Bassett, DL. A Stereoscopic Atlas of Human Anatomy. Portland: Sawyer’s Inc; 1952.Beaulieu C, Allen PS. Determinants of anisotropic water diffusion in nerves. Magn Reson Med

1994;31:394–400. [PubMed: 8208115]Bhagat YA, Beaulieu C. Diffusion anisotropy in subcortical white matter and cortical gray matter:

changes with aging and the role of CSF-suppression. J Magn Reson Imaging 2004;20:216–227.[PubMed: 15269946]

Bucur B, Madden DJ, Spaniol J, Provenzale JM, Cabeza R, White LE, Huettel SA. Age-related slowingof memory retrieval: Contributions of perceptual speed and cerebral white matter integrity. NeurobiolAging. 2007

Burgel U, Amunts K, Hoemke L, Mohlberg H, Gilsbach JM, Zilles K. White matter fiber tracts of thehuman brain: three-dimensional mapping at microscopic resolution, topography and intersubjectvariability. Neuroimage 2006;29:1092–1105. [PubMed: 16236527]

Cabeza R. Hemispheric asymmetry reduction in older adults: the HAROLD model. Psychol Aging2002;17:85–100. [PubMed: 11931290]

Catani M, Howard RJ, Pajevic S, Jones DK. Virtual in vivo interactive dissection of white matter fasciculiin the human brain. Neuroimage 2002;17:77–94. [PubMed: 12482069]

Concha L, Gross DW, Wheatley BM, Beaulieu C. Diffusion tensor imaging of time-dependent axonaland myelin degradation after corpus callosotomy in epilepsy patients. Neuroimage 2006;32:1090–1099. [PubMed: 16765064]

Corouge I, Fletcher PT, Joshi S, Gouttard S, Gerig G. Fiber tract-oriented statistics for quantitativediffusion tensor MRI analysis. Med Image Anal 2006;10:786–798. [PubMed: 16926104]

Curran EJ. A new association fiber tract in the cerebrum. With remarks on the fiber tract dissection ofstudying the brain. J Comp Neurol 1909;19:645–656.

Daselaar, SM.; Browndyke, J.; Cabeza, R. Functional neuroimaging of cognitive aging. In: Cabeza, R.;Kingstone, A., editors. Handbook of Functional Neuroimaging of Cognition. Vol. 2. MIT Press; 2006.

Daselaar, SM.; Cabeza, R. Age-related Changes in Hemispheric Organization. In: Cabeza, R.; Nyberg,L.; Park, DC., editors. Cognitive Neuroscience of Aging. Oxford: Oxford University Press; 2005.

Dubois J, Dehaene-Lambertz G, Perrin M, Mangin JF, Cointepas Y, Duchesnay E, Le Bihan D, Hertz-Pannier L. Asynchrony of the early maturation of white matter bundles in healthy infants: quantitativelandmarks revealed noninvasively by diffusion tensor imaging. Hum Brain Mapp 2008;29:14–27.[PubMed: 17318834]

Duvernoy, H. The Human Brain, Surface, Blood Supply, and Three-Dimensional Sectional Anatomy.Vienna: Springer-Verlag; 1999.

Ebeling U, von Cramon D. Topography of the uncinate fascicle and adjacent temporal fiber tracts. ActaNeurochir 1992;115:143–148.

Eluvathingal TJ, Hasan KM, Kramer L, Fletcher JM, Ewing-Cobbs L. Quantitative diffusion tensortractography of association and projection fibers in normally developing children and adolescents.Cereb Cortex 2007;17:2760–2768. [PubMed: 17307759]

Feldman ML, Peters A. Ballooning of myelin sheaths in normally aged macaques. J Neurocytol1998;27:605–614. [PubMed: 10405027]

Flechsig P. Developmental (myelogenetic) localisation of the cerebral cortex in the human subject. Lancet1901;2:1027–1029.

Gao W, Lin W, Chen Y, Gerig G, Smith JK, Jewells V, Gilmore JH. Temporal and Spatial Developmentof Axonal Maturation and Myelination of White Matter in the Developing Brain. AJNR Am JNeuroradiol. 2008

Geula C, Wu CK, Saroff D, Lorenzo A, Yuan M, Yankner BA. Aging renders the brain vulnerable toamyloid beta-protein neurotoxicity. Nat Med 1998;4:827–831. [PubMed: 9662375]

Giedd JN, Blumenthal J, Jeffries NO, Castellanos FX, Liu H, Zijdenbos A, Paus T, Evans AC, RapoportJL. Brain development during childhood and adolescence: a longitudinal MRI study. Nat Neurosci1999;2:861–863. [PubMed: 10491603]

Giorgio A, Watkins KE, Douaud G, James AC, James S, De Stefano N, Matthews PM, Smith SM,Johansen-Berg H. Changes in white matter microstructure during adolescence. Neuroimage2008;39:52–61. [PubMed: 17919933]

Davis et al.


NIH


NIH


NIH


Glenn OA, Henry RG, Berman JI, Chang PC, Miller SP, Vigneron DB, Barkovich AJ. DTI-based three-dimensional tractography detects differences in the pyramidal tracts of infants and children withcongenital hemiparesis. J Magn Reson Imaging 2003;18:641–648. [PubMed: 14635148]

Glisky EL, Polster MR, Routhieaux BC. Double dissociation between item and source memory.Neuropsychology 1995;9:229–235.

Glisky EL, Rubin SR, Davidson PS. Source memory in older adults: an encoding or retrieval problem?J Exp Psychol Learn Mem Cogn 2001;27:1131–1146. [PubMed: 11550742]

Goodlett CB, Fletcher PT, Gilmore JH, Gerig G. Group analysis of DTI fiber tract statistics withapplication to neurodevelopment. Neuroimage. 2008

Gould RL, Brown RG, Owen AM, ffytche DH, Howard RJ. fMRI BOLD response to increasing taskdifficulty during successful paired associates learning. Neuroimage 2003;20:1006–1019. [PubMed:14568470]

Grady CL, McIntosh AR, Horwitz B, Maisog JM, Ungerleider LG, Mentis MJ, Pietrini P, Schapiro MB,Haxby JV. Age-related reductions in human recognition memory due to impaired encoding. Science1995;269:218–221. [PubMed: 7618082]

Greenwood PM. The frontal aging hypothesis evaluated. J Int Neuropsychol Soc 2000;6:705–726.[PubMed: 11011517]

Grieve SM, Williams LM, Paul RH, Clark CR, Gordon E. Cognitive aging, executive function, andfractional anisotropy: a diffusion tensor MR imaging study. AJNR Am J Neuroradiol 2007;28:226–235. [PubMed: 17296985]

Gutchess AH, Welsh RC, Hedden T, Bangert A, Minear M, Liu LL, Park DC. Aging and the neuralcorrelates of successful picture encoding: frontal activations compensate for decreased medial-temporal activity. J Cogn Neurosci 2005;17:84–96. [PubMed: 15701241]

Hayakawa K, Konishi Y, Kuriyama M, Konishi K, Matsuda T. Normal brain maturation in MRI. Eur JRadiol 1991;12:208–215. [PubMed: 1855514]

Head D, Buckner RL, Shimony JS, Williams LE, Akbudak E, Conturo TE, McAvoy M, Morris JC, SnyderAZ. Differential vulnerability of anterior white matter in nondemented aging with minimalacceleration in dementia of the Alzheimer type: evidence from diffusion tensor imaging. CerebCortex 2004;14:410–423. [PubMed: 15028645]

Husain J, Juurlink BH. Oligodendroglial precursor cell susceptibility to hypoxia is related to poor abilityto cope with reactive oxygen species. Brain Res 1995;698:86–94. [PubMed: 8581507]

Jensen JH, Helpern JA, Ramani A, Lu H, Kaczynski K. Diffusional kurtosis imaging: the quantificationof non-gaussian water diffusion by means of magnetic resonance imaging. Magn Reson Med2005;53:1432–1440. [PubMed: 15906300]

Kennedy KM, Erickson KI, Rodrigue KM, Voss MW, Colcombe SJ, Kramer AF, Acker JD, Raz N. Age-related differences in regional brain volumes: A comparison of optimized voxel-based morphometryto manual volumetry. Neurobiol Aging. 2008

Kier EL, Staib LH, Davis LM, Bronen RA. MR imaging of the temporal stem: anatomic dissectiontractography of the uncinate fasciculus, inferior occipitofrontal fasciculus, and Meyer’s loop of theoptic radiation. AJNR Am J Neuroradiol 2004;25:677–691. [PubMed: 15140705]

Kinney HC, Brody BA, Kloman AS, Gilles FH. Sequence of central nervous system myelination in humaninfancy. II. Patterns of myelination in autopsied infants. J Neuropathol Exp Neurol 1988;47:217–234. [PubMed: 3367155]

Lamantia AS, Rakic P. Cytological and quantitative characteristics of four cerebral commissures in therhesus monkey. J Comp Neurol 1990;291:520–537. [PubMed: 2329189]

Landis JR, Koch GG. The measurement of observer agreement for categorical data. Biometrics1977;33:159–174. [PubMed: 843571]

Lebel C, Walker L, Leemans A, Phillips L, Beaulieu C. Microstructural maturation of the human brainfrom childhood to adulthood. Neuroimage 2008;40:1044–1055. [PubMed: 18295509]

Madden DJ, Spaniol J, Bucur B, Whiting WL. Age-related increase in top-down activation of visualfeatures. Q J Exp Psychol (Colchester) 2007;60:644–651. [PubMed: 17455072]

Madden DJ, Spaniol J, Costello MC, Bucur B, White LE, Cabeza R, Davis SW, Dennis NA, ProvenzaleJM, Huettel SA. Cerebral White Matter Integrity Mediates Adult Age Differences in CognitivePerformance. J Cogn Neurosci. in press

Davis et al.


NIH


NIH


NIH


Madden DJ, Spaniol J, Whiting WL, Bucur B, Provenzale JM, Cabeza R, White LE, Huettel SA. Adultage differences in the functional neuroanatomy of visual attention: A combined fMRI and DTI study.Neurobiol Aging. 2006

Madden DJ, Turkington TG, Provenzale JM, Denny LL, Langley LK, Hawk TC, Coleman RE. Agingand attentional guidance during visual search: functional neuroanatomy by positron emissiontomography. Psychol Aging 2002;17:24–43. [PubMed: 11931285]

Madden DJ, Whiting WL, Huettel SA, White LE, MacFall JR, Provenzale JM. Diffusion tensor imagingof adult age differences in cerebral white matter: relation to response time. Neuroimage2004;21:1174–1181. [PubMed: 15006684]

Markowitsch HJ. Which brain regions are critically involved in the retrieval of old episodic memory?Brain Res Brain Res Rev 1995;21:117–127. [PubMed: 8866670]

Meier-Ruge W, Ulrich J, Bruhlmann M, Meier E. Age-related white matter atrophy in the human brain.Ann N Y Acad Sci 1992;673:260–269. [PubMed: 1485724]

Mori S, van Zijl PC. Fiber tracking: principles and strategies - a technical review. NMR Biomed2002;15:468–480. [PubMed: 12489096]

Moscovitch, M.; Winocur, G. The neuropsychology of memory and aging. In: Craik, FIM.; Salthouse,TA., editors. The Handbook of Aging and Cognition. Hillsdale, NJ: Lawrence Erlbaum Associates;1992. p. 315-372.

O’Donnell L, Westin CF. High-dimensional white matter atlas generation and group analysis. Med ImageComput Comput Assist Interv Int Conf Med Image Comput Comput Assist Interv 2006;9:243–251.[PubMed: 17354778]

O’Sullivan M, Jones DK, Summers PE, Morris RG, Williams SC, Markus HS. Evidence for cortical“disconnection” as a mechanism of age-related cognitive decline. Neurology 2001;57:632–638.[PubMed: 11524471]

Owen AM, Morris RG, Sahakian BJ, Polkey CE, Robbins TW. Double dissociations of memory andexecutive functions in working memory tasks following frontal lobe excisions, temporal lobeexcisions or amygdalo-hippocampectomy in man. Brain 1996;119 ( Pt 5):1597–1615. [PubMed:8931583]

Paus T, Zijdenbos A, Worsley K, Collins DL, Blumenthal J, Giedd JN, Rapoport JL, Evans AC. Structuralmaturation of neural pathways in children and adolescents: in vivo study. Science 1999;283:1908–1911. [PubMed: 10082463]

Persson J, Nyberg L, Lind J, Larsson A, Nilsson LG, Ingvar M, Buckner RL. Structure-function correlatesof cognitive decline in aging. Cereb Cortex 2006;16:907–915. [PubMed: 16162855]

Peters A, Moss MB, Sethares C. Effects of aging on myelinated nerve fibers in monkey primary visualcortex. J Comp Neurol 2000;419:364–376. [PubMed: 10723011]

Peters A, Sethares C. Aging and the myelinated fibers in prefrontal cortex and corpus callosum of themonkey. J Comp Neurol 2002;442:277–291. [PubMed: 11774342]

Pfefferbaum A, Adalsteinsson E, Sullivan EV. Frontal circuitry degradation marks healthy adult aging:Evidence from diffusion tensor imaging. Neuroimage 2005;26:891–899. [PubMed: 15955499]

Price G, Cercignani M, Parker GJ, Altmann DR, Barnes TR, Barker GJ, Joyce EM, Ron MA. Whitematter tracts in first-episode psychosis: a DTI tractography study of the uncinate fasciculus.Neuroimage 2008;39:949–955. [PubMed: 17988894]

Rabbitt P, Lowe C. Patterns of cognitive ageing. Psychol Res 2000;63:308–316. [PubMed: 11004884]Raz, N. Aging of the Brain and Its Impact on Cognitive Performance: Integration of Structural and

Functional Findings. In: Craik, FIM.; Salthouse, TA., editors. The Handbook of Aging and Cognition.London: Lawrence Erlbaum; 2000.

Raz, N. The Aging Brain Observed in Vivo: Differential Changes and Their Modifiers. In: Cabeza, R.;Nyberg, L.; Park, DC., editors. Cognitive Neuroscience of Aging. Oxford University Press; 2005. p.19-57.

Raz N, Gunning FM, Head D, Dupuis JH, McQuain J, Briggs SD, Loken WJ, Thornton AE, Acker JD.Selective aging of the human cerebral cortex observed in vivo: differential vulnerability of theprefrontal gray matter. Cereb Cortex 1997;7:268–282. [PubMed: 9143446]

Davis et al.


NIH


NIH


NIH


Raz N, Lindenberger U, Rodrigue KM, Kennedy KM, Head D, Williamson A, Dahle C, Gerstorf D,Acker JD. Regional brain changes in aging healthy adults: general trends, individual differences andmodifiers. Cereb Cortex 2005;15:1676–1689. [PubMed: 15703252]

Resnick SM, Pham DL, Kraut MA, Zonderman AB, Davatzikos C. Longitudinal magnetic resonanceimaging studies of older adults: a shrinking brain. J Neurosci 2003;23:3295–3301. [PubMed:12716936]

Robbins TW, James M, Owen AM, Sahakian BJ, Lawrence AD, McInnes L, Rabbitt PM. A study ofperformance on tests from the CANTAB battery sensitive to frontal lobe dysfunction in a large sampleof normal volunteers: implications for theories of executive functioning and cognitive aging.Cambridge Neuropsychological Test Automated Battery. J Int Neuropsychol Soc 1998;4:474–490.[PubMed: 9745237]

Ronen I, Ugurbil K, Kim D. How Does DWI Correlate With White Matter Structures? MAGNETICRESONANCE IN MEDICINE 2005;54:317. [PubMed: 16032693]

Sakuma H, Nomura Y, Takeda K, Tagami T, Nakagawa T, Tamagawa Y, Ishii Y, Tsukamoto T. Adultand neonatal human brain: diffusional anisotropy and myelination with diffusion-weighted MRimaging. Radiology 1991;180:229–233. [PubMed: 2052700]

Salat DH, Kaye JA, Janowsky JS. Prefrontal gray and white matter volumes in healthy aging andAlzheimer disease. Arch Neurol 1999;56:338–344. [PubMed: 10190825]

Salat DH, Tuch DS, Hevelone ND, Fischl B, Corkin S, Rosas HD, Dale AM. Age-related changes inprefrontal white matter measured by diffusion tensor imaging. Ann N Y Acad Sci 2005;1064:37–49.[PubMed: 16394146]

Scahill RI, Frost C, Jenkins R, Whitwell JL, Rossor MN, Fox NC. A longitudinal study of brain volumechanges in normal aging using serial registered magnetic resonance imaging. Arch Neurol2003;60:989–994. [PubMed: 12873856]

Schmierer K, Scaravilli F, Altmann DR, Barker GJ, Miller DH. Magnetization transfer ratio and myelinin postmortem multiple sclerosis brain. Ann Neurol 2004;56:407–415. [PubMed: 15349868]

Smith SM, Jenkinson M, Johansen-Berg H, Rueckert D, Nichols TE, Mackay CE, Watkins KE, CiccarelliO, Cader MZ, Matthews PM, Behrens TE. Tract-based spatial statistics: voxelwise analysis of multi-subject diffusion data. Neuroimage 2006;31:1487–1505. [PubMed: 16624579]

Smith SM, Jenkinson M, Woolrich MW, Beckmann CF, Behrens TE, Johansen-Berg H, Bannister PR,De Luca M, Drobnjak I, Flitney DE, Niazy RK, Saunders J, Vickers J, Zhang Y, De Stefano N, BradyJM, Matthews PM. Advances in functional and structural MR image analysis and implementation asFSL. Neuroimage 2004;23(Suppl 1):S208–219. [PubMed: 15501092]

Smith SM, Zhang Y, Jenkinson M, Chen J, Matthews PM, Federico A, De Stefano N. Accurate, robust,and automated longitudinal and cross-sectional brain change analysis. Neuroimage 2002;17:479–489. [PubMed: 12482100]

Snook L, Paulson LA, Roy D, Phillips L, Beaulieu C. Diffusion tensor imaging of neurodevelopment inchildren and young adults. Neuroimage 2005;26:1164–1173. [PubMed: 15961051]

Song SK, Sun SW, Ju WK, Lin SJ, Cross AH, Neufeld AH. Diffusion tensor imaging detects anddifferentiates axon and myelin degeneration in mouse optic nerve after retinal ischemia. Neuroimage2003;20:1714–1722. [PubMed: 14642481]

Song SK, Sun SW, Ramsbottom MJ, Chang C, Russell J, Cross AH. Dysmyelination revealed throughMRI as increased radial (but unchanged axial) diffusion of water. Neuroimage 2002;17:1429–1436.[PubMed: 12414282]

Song SK, Yoshino J, Le TQ, Lin SJ, Sun SW, Cross AH, Armstrong RC. Demyelination increases radialdiffusivity in corpus callosum of mouse brain. Neuroimage 2005;26:132–140. [PubMed: 15862213]

Sowell ER, Thompson PM, Leonard CM, Welcome SE, Kan E, Toga AW. Longitudinal mapping ofcortical thickness and brain growth in normal children. J Neurosci 2004;24:8223–8231. [PubMed:15385605]

Sowell ER, Trauner DA, Gamst A, Jernigan TL. Development of cortical and subcortical brain structuresin childhood and adolescence: a structural MRI study. Dev Med Child Neurol 2002;44:4–16.[PubMed: 11811649]

Davis et al.


NIH


NIH


NIH


Sullivan EV, Adalsteinsson E, Hedehus M, Ju C, Moseley M, Lim KO, Pfefferbaum A. Equivalentdisruption of regional white matter microstructure in ageing healthy men and women. Neuroreport2001;12:99–104. [PubMed: 11201100]

Sullivan EV, Adalsteinsson E, Pfefferbaum A. Selective age-related degradation of anterior callosal fiberbundles quantified in vivo with fiber tracking. Cereb Cortex 2006;16:1030–1039. [PubMed:16207932]

Sullivan EV, Pfefferbaum A. Diffusion tensor imaging and aging. Neurosci Biobehav Rev 2006;30:749–761. [PubMed: 16887187]

Sullivan EV, Rohlfing T, Pfefferbaum A. Quantitative fiber tracking of lateral and interhemispheric whitematter systems in normal aging: Relations to timed performance. Neurobiol Aging. 2008

Sun SW, Liang HF, Schmidt RE, Cross AH, Song SK. Selective vulnerability of cerebral white matterin a murine model of multiple sclerosis detected using diffusion tensor imaging. Neurobiol Dis2007;28:30–38. [PubMed: 17683944]

Tang Y, Nyengaard JR. A stereological method for estimating the total length and size of myelin fibersin human brain white matter. J Neurosci Methods 1997;73:193–200. [PubMed: 9196291]

Thomalla G, Glauche V, Koch MA, Beaulieu C, Weiller C, Rother J. Diffusion tensor imaging detectsearly Wallerian degeneration of the pyramidal tract after ischemic stroke. Neuroimage2004;22:1767–1774. [PubMed: 15275932]

Trip SA, Wheeler-Kingshott C, Jones SJ, Li WY, Barker GJ, Thompson AJ, Plant GT, Miller DH. Opticnerve diffusion tensor imaging in optic neuritis. Neuroimage 2006;30:498–505. [PubMed: 16242968]

Vernooij MW, de Groot M, van der Lugt A, Ikram MA, Krestin GP, Hofman A, Niessen WJ, BretelerMM. White matter atrophy and lesion formation explain the loss of structural integrity of white matterin aging. Neuroimage 2008;43:470–477. [PubMed: 18755279]

Vogt BA, Finch DM, Olson CR. Functional heterogeneity in cingulate cortex: the anterior executive andposterior evaluative regions. Cereb Cortex 1992;2:435–443. [PubMed: 1477524]

Wakana S, Caprihan A, Panzenboeck MM, Fallon JH, Perry M, Gollub RL, Hua K, Zhang J, Jiang H,Dubey P, Blitz A, van Zijl P, Mori S. Reproducibility of quantitative tractography methods appliedto cerebral white matter. Neuroimage 2007;36:630–644. [PubMed: 17481925]

Wakana S, Jiang H, Nagae-Poetscher LM, van Zijl PC, Mori S. Fiber tract-based atlas of human whitematter anatomy. Radiology 2004;230:77–87. [PubMed: 14645885]

West RL. An application of prefrontal cortex function theory to cognitive aging. Psychol Bull1996;120:272–292. [PubMed: 8831298]

Wozniak JR, Lim KO. Advances in white matter imaging: a review of in vivo magnetic resonancemethodologies and their applicability to the study of development and aging. Neurosci Biobehav Rev2006;30:762–774. [PubMed: 16890990]

Xu D, Mori S, Solaiyappan M, van Zijl PC, Davatzikos C. A framework for callosal fiber distributionanalysis. Neuroimage 2002;17:1131–1143. [PubMed: 12414255]

Yakovlev, PI.; Lecours, AR. Regional Development of the Brain in Early Life. Minkowski, A., editor.Boston: Blackwell Scientific; 1967.

Yasmin H, Nakata Y, Aoki S, Abe O, Sato N, Nemoto K, Arima K, Furuta N, Uno M, Hirai S, MasutaniY, Ohtomo K. Diffusion abnormalities of the uncinate fasciculus in Alzheimer’s disease: diffusiontensor tract-specific analysis using a new method to measure the core of the tract. Neuroradiology.2008

Davis et al.


NIH


NIH


NIH


Figure 1.Methodology for assessing diffusion metrics along fiber tract lengths. (A) Seed (red) and target(green) regions were drawn on the b0 image in order to identify white matter tracts; in thisexample an axial view of the anterior cortex depicts a seed region around the midline and leftand right ROIs in frontal cortex are used to elucidate fiber tracts passing through the genu ofthe corpus callosum; (B) identified extracted fiber tracts with outliers removed; (C) mean fiberswere constructed by averaging the spatial characteristics of all fibers within a tract. (D) Linearlyadjusted mean differences were assed via ANOVA and F-statistics were transposed back ontothe mean tract for visualization.

Davis et al.


NIH


NIH


NIH


Figure 2.Age-related differences in FA, RD, and AD for tracts of interest. FA differences are shownalong a tract; FA data were smoothed with a 6mm kernel and displayed upon mean tracts froma representative subject for the genu, UFs, CBs, ILFs, and splenium. RD and AD differenceswithin these target sections are displayed in graphs corresponding to specific long associationtracts. RD = radial diffusivity; AD = axial diffusivity; CB = cingulum bundle; UF = uncinatefasciculus; ILF = inferior longitudinal fasciculus. Metric dimensions: AD, RD: 10−3 mm2/s; *= P < 0.01; ** = P < 0.005; *** = P < 0.001.

Davis et al.


NIH


NIH


NIH


Figure 3.Age-related differences in AD (A) and RD (B) represented upon mean tracts from arepresentative subject, demonstrating greater age-related effects for RD than AD. Color scalesfor RD have been reversed such that OA > YA differences are reflected by warmer colors. RD= radial diffusivity; AD = axial diffusivity.

Davis et al.


NIH


NIH


NIH


Figure 4.Trend analysis of the CB and frontal UF. Trend lines for both linear (solid) and quadratic(dashed) trends are represented over the y-axis component (Talairach coordinates) of each tractusing the effect size (partial eta) based on a pairwise analysis at points along a tract. F statisticsfor best fit for both linear and quadratic terms are displayed below the data. Linear trends aresignificant for the left and right CB, and the left and right (frontal) UF. The vertical dashedline approximates the border between frontal and parietal cortex, according to the anteriorborder precentral sulcus. CB = cingulum bundle; UF = uncinate fasciculus; YA = young adult;OA = older adult.

Davis et al.


NIH


NIH


NIH


Figure 5.Representative correlation between diffusivity measures extracted from the right uncinatefasciculus and performance on the Intra-Extra Dimensional Set Shift (IED) task. Scatterplotsshow significant relationships between IED performance and FA and RD, but not AD. Metricdimensions: FA: dimensionless; AD, RD: 10−3 mm2/s. * = P < 0.05; *** = P < 0.005.

Davis et al.


NIH


NIH


NIH


NIH


NIH


NIH


Davis et al.

Table 1

Participant characteristics and neuropsychological data.

YA (n = 20) M (SD) OA (n = 20) M (SD)

Age 20.04 (2.85) 68.89 (7.29)Years of Education*** 13.59 (1.70) 17.47 (1.47)Mini-Mental State Examination n/a 28.3 (1.5)Executive Tasks Spatial Working Memory Between errors** 7.40 (9.97) 35.40 (24.35) Strategy* 25.00 (5.17) 32.35 (6.98) Spatial Span ** 7.80 (1.01) 5.65 (1.31) Rapid Visual Information Processing** 0.96 (0.03) 0.91 (0.06) Intra-Extra Dimensional Set Shifting Stages completed 9.00 (0.00) 8.75 (0.64) Total errors** 10.35 (4.32) 20.70 (12.72)Memory Tasks Paired Associate Learning* 96.27 (4.14) 92.08 (4.86) Pattern Recognition Memory** 2.60 (2.58) 14.20 (8.11)Speed Tasks Movement time** 316.83 (40.97) 402.82 (72.78) Reaction time** 298.91 (39.05) 376.09 (52.53)

Note: YA, younger adults; OA, older adults; M, mean; SD, standard deviation. Age-related differences indicated by

*= P < 0.05;

**= P < 0.005;

***= P < 0.001.


NIH


NIH


NIH


Davis et al.

Table 2

Fiber Tracking heuristics.

Fiber Tract Seed ROI Target ROI Midpoint IRR

Genu anterior CC frontopolar cortex midline 0.938Splenium posterior CC occipital cortex midline 0.927Uncinate fasciculus temporal pole inferior frontal cortex frontotemporal border L: 0.942; R: 0.836Cingulum bundle posterior cingulate cortex anterior cingulate cortex level of anterior commissure L: 0.881; R:0.824Inferior longitudinal fascicus anterior temporal cortex periventricular occipital cortex occipitotemporal fissure L: 0.913; R:0.892

Note: ROI, region of interest; IRR, intra-rater reliability, CC, corpus callosum


NIH


NIH


NIH


Davis et al. Ta

ble

3

Age

diff

eren

ces i

n Fr

actio

nal A

niso

tropy

(FA

), R

adia

l Diff

usiv

ity (R

D),

and

Axi

al D

iffus

ivity

(AD

).

FAR

DA

D

YA

M (S

D)

OA

M (S

D)

η pY

A M

(SD

)O

A M

(SD

)η p

YA

M (S

D)

OA

M (S

D)

η p

Gen

u0.

55 (0

.04)

0.50

(0.0

4)**

*0.

330.

49 (0

.04)

0.54

(0.0

6)**

0.22

1.27

(0.0

5)1.

24 (0

.06)

0.09

Sple

nium

0.67

(0.0

4)0.

63 (0

.04)

*0.

150.

54 (0

.12)

0.59

(0.1

1)0.

051.

78 (0

.19)

1.75

(0.1

7)0.

01L

eft C

B0.

58 (0

.04)

0.56

(0.0

4)*

0.17

0.46

(0.0

5)0.

52 (0

.04)

**0.

271.

26 (0

.09)

1.25

(0.0

9)0.

00R

ight

CB

0.55

(0.0

2)0.

51 (0

.04)

**0.

180.

47 (0

.03)

0.51

(0.0

3)**

*0.

401.

26 (0

.08)

1.23

(0.0

7)0.

01L

eft U

F‡0.

51 (0

.05)

0.46

(0.0

4)**

0.21

0.51

(0.0

6)0.

55 (0

.07)

0.10

1.23

(0.1

6)1.

17 (0

.08)

0.04

Rig

ht U

F†0.

51 (0

.04)

0.47

(0.0

3)**

0.27

0.48

(0.0

3)0.

54 (0

.07)

**0.

241.

19 (0

.04)

1.19

(0.0

6)0.

00L

eft I

LF

0.57

(0.0

4)0.

53 (0

.03)

**0.

210.

48 (0

.03)

0.55

(0.0

6)**

*0.

351.

31 (0

.07)

1.36

(0.0

7)0.

10R

ight

ILF

0.58

(0.0

4)0.

54 (0

.03)

***

0.25

0.46

(0.0

2)0.

50 (0

.05)

**0.

251.

29 (0

.06)

1.31

(0.0

6)0.

03

Not

e: M

etric

dim

ensi

ons:

FA

: dim

ensi

onle

ss; A

D, R

D: 1

0−3

mm

2 /s;

* = P

< 0.

01;

**=

P <

0.00

5;

*** =

P <

0.00

1.

YA

= y

oung

adu

lt; O

A =

old

er a

dult;

M =

mea

n; S

D =

stan

dard

dev

iatio

n; η

p =

parti

al e

ta (e

ffec

t siz

e).

† = N

= 1

9 Y

As,

20 O

As;

‡ = N

= 2

0 Y

As,

19 O

As.


NIH


NIH


NIH


Davis et al.

Table 4

Relationships between DTI metrics and cognitive performance in older adults.

Pearson’s r

tract cognitive test FA RD AD

Genu Spatial Span 0.57** −0.41* −0.18Right UF Spatial Working Memory‡ 0.57** −0.37 0.18

Spatial Working Memory – strategy 0.52** −0.40* 0.03Intra-Extra Dimensional Set Shifting‡ 0.63*** −0.38* 0.21

L CB Pattern Recognition Memory 0.38* −0.37* 0.23Paired Associate Learning‡ 0.47* −0.45* −0.06

L ILF Paired Associate Learning‡ 0.43* −0.42* −0.32Splenium Pattern Recognition Memory 0.44* −0.53** −0.46*

‡Notes: = Task scores based on number of errors have directions of Pearson’s r flipped.

*= P < 0.05;

**= P < 0.01;

***= P < 0.005


author manuscript nih public access norbou g. buchler leonard … · introduction post-mortem and...

Documents